A natural question to ask in further diagnosing Europe’s comparatively low BERD is whether it persists at the level of individual industries. Indeed, Europe’s low overall BERD could re‡ect (i) low R&D intensity –de…ned as BERD relative to value added – in most or all industries (R&D intensity e¤ect), (ii) an industry composition e¤ect whereby Europe might be specialised on industries relying less on formalized R&D, or (iii) a combination of (i) and (ii)61.
In answering this question, it is useful to start by showing which industries spend most on R&D. Figure 3.3 gives this information for the three economic zones of the triad. Three main insights emerge. First, three broad manufacturing-industry groups account for the brunt of R&D: Chemicals and pharmaceuticals (ISIC 24), Transport equipment (ISIC 34 and 35) and ICT and other non-transport equipment (ISIC 29 to 33). These industries make up three quarters of aggregate BERD in the EU and even 80 percent in Japan although they produce less than one-tenth of GDP62. Second,
within these three leading industry groups, Japan’s R&D is more concentrated on ICT equipment than R&D in the EU and the US while Europe has a stronger focus on Transport equipment. Third, outside the three leading manufacturing industry groups, the US records a signi…cant share of BERD – almost one third – in services whereas Japan spends a lot on other manufacturing.
61This subsection draws on and updates Uppenberg (2009b).
62Because of their high R&D intensity, the individual industries in the three broad groups are all
labelled as either high-technology or medium-to-high-technology in the OECD’s classi…cation of tech- nology intensities in manufacturing while the remaining manufacturing industries are “low-technology” or “medium-to-low technology”. See Table A1 of Danguy et al. (2009) for an overview of all individual manufacturing industries.
Business R&D expenditure and capital in Europe 60
Figure 3.3: BERD by industry groups, EU, 2005
However, a strong caveat must be put on international comparisons of BERD at the industry level. According to international conventions, R&D statistics should allocate each R&D activity to the targeted product …eld (e.g. a new computer) rather than the main activity (measured by turnover) of the R&D-performing company. Moreover, R&D activities by specialised R&D service …rms (ISIC 73) should be allocated to the industries purchasing these services. Countries di¤er as to whether they follow these conventions. This matters for the reported industry breakdown of BERD (see also appendix 3.2).
We therefore distinguish between three groups of countries by decreasing degree of comparability when comparing individual EU countries and their industry-level R&D data. Country group 1 comprises countries that follow the product …eld approach in collecting R&D data. These are Belgium, Finland, France, Sweden and the UK. We
Business R&D expenditure and capital in Europe 61
also include Germany and the Netherlands which, albeit following the main-activity approach, break down the R&D expenditure of their biggest R&D-performing compa- nies by product …eld. The other countries collect BERD by companies’main activity. Country group 2 comprises countries that reallocate part or all of the BERD by R&D service …rms to the consuming industries, most often located in manufacturing. All other countries are in country group 3. The bulk of BERD in the EU is done in group-1 countries whereas the US and Japan fall into group 3.
With that caveat in mind, we now look at industry-level R&D intensities and in- dustry composition in order to understand what accounts for Europe’s gap in overall BERD. In doing so, we focus on the three most R&D-intensive manufacturing industry groups that account for some three quarters of total BERD. Figure 3.4 shows that the lower overall R&D intensity in the EU compared with the US and Japan applies to all three industry groups. The chemical and pharmaceutical industry of Japan spent 23 percent of its value added on R&D in 2005, compared with 18 percent and 13 percent for their US and EU counterparts, respectively. Europe’s gap is even larger in ICT and other non-transport equipment industries. By contrast, it is small in Transport equip- ment where R&D intensities are broadly the same throughout the triad at between 15 and 18 percent. The …rst conclusion therefore is that the R&D intensity e¤ect is at work in key industries. Arguably, this accounts for a good part of Europe’s gap in overall R&D expenditure vis-à-vis the US and Japan.
Nevertheless, this does not necessarily mean that lower R&D intensity accounts for all of the gap because di¤erences in specialization might matter, too. This would be the case if the output of technology-intensive manufacturing industries were smaller in the EU compared with the US and Japan. Figure 3.5 shows the share of each industry group’s value added in aggregate value added for the EU, the US and Japan. When measured at current prices – as is done in the left half of …gure 3.5 – technology- intensive manufacturing contributed 8 percent to aggregate value added in the EU, more than in the US (6 percent) but less than in Japan (10 percent). Thus, it seems that the EU is more specialized in technology-intensive manufacturing production than the US and, hence, that the gap vis-à-vis the US is entirely due to lower industry R&D intensities.
Business R&D expenditure and capital in Europe 62
Figure 3.4: R&D intensity in technology-intensive industries in the triad (percent of industry real value added), 2005
However, things look di¤erent when basing the analysis on real value added. The right half of …gure 3.5 depicts each industry’s contribution to real value added, i.e. value added in prices of 1995. From this perspective, the EU is less specialized than the US on technology-intensive manufacturing (share of 9 percent compared with 1212 percent) while Japan continues to be most specialized (16 percent). The di¤erence between real and nominal shares stems from ICT and other non-transport industries and is particularly pronounced in the US and Japan but small for the EU. This is because within this broad industry group, the US, and even more so Japan, are specialized on ICT-equipment production where prices decline much faster than in other industries such as machine-tools and optical instruments. Since these price declines are themselves to a large extent technology-driven and, hence, dependent on R&D, it makes sense to assess the industry’s contribution to the level of GDP on real value added63. Closer inspection of the right half of …gure 3.5 suggests that the share in real value added of ICT and other non-transport equipment is signi…cantly smaller in the EU than in the US and in Japan, pointing to an industry composition e¤ect alongside the R&D 63By contrast, nominal value added is more appropriate to assess the resource cost of R&D as
Business R&D expenditure and capital in Europe 63
intensity e¤ect mentioned above.
Figure 3.5: Industry composition of value added in the triad, 2005
0% 5% 10% 15% 20% 25% 30% 35% 40% 45% 50% EU-27 US JP EU-27 US JP
Nominal GVA Real GVA (prices of 1995)
Chemicals & Pharmaceuticals Transport equipment ICT and other non-transport equipment Other manufacturing
Business services Other industries (up to 100%)
Source: Eurostat, own calculations
Notes: In the right half of the figure, the share of each industry is defined as the ratio of the industry's real value added to total-economy real value added. The share of "All other industries" equals the difference between 100 percent and the sum of the five industry groups shown.
All in all, this section has shown that R&D expenditure in the EU lags behind that in the US and Japan, which is attributable to the business sector rather than the government sector. There has been no sign of the EU catching-up with the other areas of the triad over the past 15 years. BERD is heavily concentrated on three technology-intensive manufacturing industry groups: Transport equipment, ICT and other equipment, and Chemicals and Pharmaceuticals. The lower R&D intensities in the latter two and the modest size of Europe’s ICT-producing industries account for most of the shortfall in overall BERD.
Business R&D expenditure and capital in Europe 64